Alzheimer's disease (AD) is one of the greatest public health challenges in the United States. There is no cure for AD, but pharmaceutical companies and academia are investigating several disease-modifying medicines that target early stages of AD neuropathology, before the damage to the brain is irreparable. However, this re- search is impeded by the enormous costs of conducting AD clinical trials. These costs are high because it is difficult to identify individuals who have early symptomatic or presymptomatic AD, as well as because AD develops slowly and it takes a very long time to discover whether a treatment is effective. The goal of this project is to develop novel neuroimaging biomarkers that can serve as surrogate measures of brain degeneration in AD. This study will build on the success of the NIH/NIA Alzheimer<s Disease Neuroimaging Initiative (ADNI), which has investigated the effectiveness of brain imaging as a biomarker by imaging the brains of hundreds of subjects semiannually over the course of three years. One of the key early conclusions of this study is that the volume of the hippocampus, as estimated from brain MRI, is a highly sensitive bio- marker that can likely increase the power of treatment effect detection by an order of magnitude over traditional psychometrics. Our project will build on this finding by developing novel biomarkers that examine hippocampal anatomy with greater precision and take into account the heterogeneity of this complex brain structure. Our project differs from ADNI (a) because it uses an MRI protocol that specifically targets the hippocampus and surrounding structures, (b) because it leverages postmortem MRI and histology data for the analysis of in vivo imaging data; and (c) because it will yield more detailed measurements describing the volume, thickness and shape of the individual subfields of the hippocampus and related structures such as the entorhinal cortex. The specific aims of this project are (1) to build a detailed three-dimensional computational atlas of the human hippocampus and entorhinal cortex using a combination of ultra high-resolution 9.4 Tesla MRI of autopsy tissue samples and histology; (2) to develop algorithms and software that would leverage this atlas for automatic detection of the subfields of the hippocampus and entorhinal cortex in in vivo MRI acquired with a T2-weighted protocol that targets this region; (3) to compare the effectiveness of 3 Tesla and 7 Tesla MRI scanners for imaging the medial temporal lobe and deriving biomarkers; and (4) to assess the sensitivity and specificity of the novel biomarkers for progression detection and cohort stratification in AD using imaging data from healthy elderly, individuals with mild cognitive impairment, AD patients, and patients with frontotemporal dementia. PUBLIC HEALTH RELEVANCE: Alzheimer's disease is one of the greatest challenges public health challenges in the United States. This project aims to reduce the cost of conducting a clinical trial in Alzheimer's disease by making it easier to detect the effects of treatment in a relatively short time window and with fewer participants. If successful, this project may result in more clinical trials being launched and a greater likelihood of finding a cure for Alzheimer's disease.